Abstract:

Recently, machine Learning algorithms are widely used in many fields. Especially, they are really good to create prediction models for problems which are not easy to solve with conventional programming techniques. Although, there are many different kinds of machine learning algorithms, results of applications are varying depend on type of data and correlation of information. In this study, different machine learning algorithms have been used for appliance recognition. The measurement data of Appliance Consumption Signatures database and some derivative values have been used for training and testing. Additionally, a data pre-processing technique and its effects on results have been presented. Filtering corrupted data and removing uncertain measurement value has improved the quality of machine learning. Combination of machine learning algorithms is best way to work with uncertain values. Different feature extraction methods and data pre-processing techniques are crucial in machine learning. Therefore, this study aims to develop a high accurate appliance recognition technique by combining grey relational analysis and an ensemble classification method. The results of this new method have been presented comparatively to show the improvement for itself and previous studies that uses the same database.

Keywords: